基于注意力机制的TCN-BiLSTM船舶轨迹预测  

Prediction for TCN-BiLSTM Ship Trajectory Based on Attention Mechanism

作　　者：郭逸婕 张君毅[1,2] 王鹏 GUO Yijie;ZHANG Junyi;WANG Peng(The 54th Research Institution of CETC,Shijiazhuang 050081,China;Hebei Province Key Laboratory of Electromagnetic Specturm Cognition and Control,Shijiazhuang 050081,China)

机构地区：[1]中国电子科技集团公司第54研究所,石家庄050081 [2]河北省电磁频谱认知与管控重点实验室,石家庄050081

出　　处：《计算机测量与控制》2024年第1期30-36,共7页Computer Measurement &Control

基　　金：国家自然科学基金(U19B2028);第六届中国科学青年人才托举工程项目(2020QNRC001)。

摘　　要：针对现有船舶轨迹预测模型预测准确度低的问题,提出一种基于注意力机制的时域卷积网络和双向长短时记忆网络结合的船舶轨迹预测模型;首先搭建TCN网络提取船舶轨迹的序列特征,之后将注意力机制引入网络调整不同属性特征的权值,突出对轨迹预测影响更大的特征,最后搭建Bi-LSTM网络学习轨迹序列的前后状况来提取序列中更多的信息,实现对船舶未来轨迹的预测;通过实际船舶AIS数据对网络进行训练与测试实验,实验结果表明,TCN-ABiLSTM模型相比LSTM、Bi-LSTM和BiLSTM-Attention模型船舶轨迹预测精度更高,拟合程度更好,验证了所设计的TCN-ABiLSTM模型在船舶轨迹预测方面的的有效性和实用性。To address the problem of low prediction accuracy in existing ship trajectory prediction model,a ship trajectory prediction model based on attention mechanism time-domain convolutional network and bidirectional long-short memory network is proposed Firstly,the temporal convolutional network(TCN)network is constructed to extract the sequence features of ship trajectories.Then,attention mechanism is introduced into the network to adjust the weights of different attribute features,highlighting greater influence on the trajectory prediction.Finally,the bi-directional long short-term memory(Bi-LSTM)network is constructed to learn the pre and post situation of trajectory sequences to extract more information from the sequences,achieving the prediction of future ship trajectories;Training and testing experiments are conducted on the network by using actual ship automatic identification system(AIS)data.The experimental results show that compared to the LSTM,Bi-LSTM and BiLSTM-Attention models,the TCN-ABiLSTM model has higher accuracy and better fit in predicting ship trajectories.which verifyes the effectiveness and practicality of the proposed TCN-ABiLSTM model in predicting ship trajectories.

关 键 词：轨迹预测 时域卷积网络 长短时记忆网络 注意力机制 AIS 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]